Osayuki Ehigie

Senior Native Android Engineer·AI Features With Evaluation·New York City


347-935-2386·oethe68@gmail.com·osayukiehigie.com·linkedin.com/in/osayuki-e-356254b1

Six years building and shipping native Android in Kotlin. Three apps live on Google Play, including one I architected and shipped as founder of my own company. I also build AI features into mobile products and prove they work before they ship. Fixed scope, fixed price, so you know what you are getting and what it costs before we start.

AI Feature, Proven

$12,000 to $25,000·4 to 6 weeks

One production AI feature integrated into your existing app, delivered with an evaluation harness that proves it works. Feature design and implementation, LLM API or on-device ML integration, a defined test set with correct answers set in advance, agreement scoring, iteration, regression tracking, and a written evaluation report you own.

Most people shipping AI features cannot tell you whether the output is correct. I build the feature and I build the proof.

Android Modernize and Comply

$8,000 to $20,000·3 to 5 weeks

Target API 36 migration, Google Play Billing Library 8 migration, Java to Kotlin conversion of core modules, XML Views to Jetpack Compose on primary screens, dependency and Gradle cleanup, and a build that passes Play review.

As of August 31, 2026, apps that do not target API 36 stop reaching new users on Google Play. Extensions run to November 1. If you missed the date, that is the window we are working in.

App Rescue

$500 audit, credited toward the build·3-day turnaround

Your developer disappeared, the agency stopped answering, or the build stalled. Full source review, a written assessment of code quality, architecture and store-readiness, what is broken and what is missing, and a fixed-price quote to finish it.

Three days to know where you stand. Fixed price, so you are not billed by the hour by someone you have not worked with yet.

Three Android apps live on Google Play

Two shipped at Parabit Systems for enterprise banking and manufacturing clients, both in daily production use. Improved app performance 40% and increased Bluetooth pairing success 30% through targeted debugging. The third is my own company’s product, architected and shipped July 2026.

Published AI evaluation writeup

github.com/OEThe11/subscout-evaluation — a 26-decision test set with correct answers defined in advance, model agreement raised from 53.8% to 73.1% across three iterations, including a change that broke three unrelated categories and was reverted. The methodology and the failures are both public.

App Store review resolution

My iOS build was rejected by Apple under a category restriction most developers do not know exists. I diagnosed it, repositioned the product, and resubmitted. I know what gets apps rejected and how to get them through.

Book a 20-minute call and leave with a scope and a number.

No pitch deck, no discovery retainer. Tell me what you are trying to ship and I will tell you what it costs and how long it takes.

oethe68@gmail.com·347-935-2386